CNS-targeted glucocorticoid reduces pathology in mouse model of amyotrophic lateral sclerosis

Abstract

Background: Hallmarks of CNS inflammation, including microglial and astrocyte activation, are prominent features in post-mortem tissue from amyotrophic lateral sclerosis (ALS) patients and in mice overexpressing mutant superoxide dismutase-1 (SOD1G93A). Administration of non-targeted glucocorticoids does not significantly alter disease progression, but this may reflect poor CNS delivery. Here, we sought to discover whether CNS-targeted, liposomal encapsulated glucocorticoid would inhibit the CNS inflammatory response and reduce motor neuron loss. SOD1G93A mice were treated with saline, free methylprednisolone (MP, 10 mg/kg/week) or glutathione PEGylated liposomal MP (2B3-201, 10 mg/kg/week) and compared to saline treated wild-type animals. Animals were treated weekly with intravenous injections for 9 weeks from 60 days of age. Weights and motor performance were monitored during this period. At the end of the experimental period (116 days) mice were imaged using T2-weighted MRI for brainstem pathology; brain and spinal cord tissue were then collected for histological analysis.

Results: All SOD1G93A groups showed a significant decrease in motor performance, compared to baseline, from ~100 days. SOD1G93A animals showed a significant increase in signal intensity on T2 weighted MR images, which may reflect the combination of neuronal vacuolation and glial activation in these motor nuclei. Treatment with 2B3-201, but not free MP, significantly reduced T2 hyperintensity observed in SOD1G93A mice. Compared to saline-treated and free-MP-treated SOD1G93A mice, those animals given 2B3-201 displayed significantly improved histopathological outcomes in brainstem motor nuclei, which included reduced gliosis and neuronal loss.

Conclusions: In contrast to previous reports that employed free steroid preparations, CNS-targeted anti-inflammatory agent 2B3-201 (liposomal methylprednisolone) has therapeutic potential, reducing brainstem pathology in the SOD1G93A mouse model of ALS. 2B3-201 reduced neuronal loss and vacuolation in brainstem nuclei, and reduced activation preferentially in astrocytes compared with microglia. These data also suggest that other previously ineffective therapies could be of therapeutic value if delivered specifically to the CNS.

Figures

Figure 1

Disease progression assessed by weight and motor behaviour. Weight data (unadjusted – A; normalized at 60 days – B) shows a gradual increase in weight in all groups, with no significant difference between genotypes or treatment groups at any time point. Rotarod scores (C) were higher at all time points for WT compared with all SOD1G93A treatment groups, and all SOD1G93A groups showed a typical decline over time, with no significant difference between any of the treatment groups. There was no significant difference between groups for weight data with area-under-the-curve analysis (D); SOD1G93A 2B3-201 had significantly greater area under the curve (increased motor scores) compared with free MP, but there was no difference when compared with SOD1G93A saline. Data are presented as mean ± standard error. *** p<0.001.

Figure 2

Changes inT2MR signal intensity in cranial motor nuclei following steroid treatment. An increase in signal is present in the trigeminal (V), facial (VII) and hypoglossal (XII) nuclei in SOD1G93A saline mice (D-F) compared to WT animals (A-C). A comparable signal increase in seen for SOD1G93A mice treated with free MP (G-I), but in comparison to saline and free MP treated SOD1G93A mice, those treated with 2B3-201 (J-L) show a markedly reduced signal intensity in these three motor nuclei. These data are presented graphically for the trigeminal, facial and hypoglossal nuclei in M-O respectively. * p < 0.05; ** p<0.01; *** p<0.001.

Figure 3

MutantSOD1-induced vacuolation is ameliorated by steroid treatment. WT tissue shows staining of large motor neurons (black arrows) in the cranial motor nuclei (A, E, I) and spinal cord (M), with no structural abnormalities. In contrast, SOD1G93A saline (B, F, J, N) and SOD1G93A free MP (C, G, K, O) show a reduction in large motor neurons, with more morphologically abnormal neurons (white arrows). There is also substantial vacuolation. SOD1G93A mice treated with 2B3-201 (D, H, L, P) have reduced vacuolation in the cranial motor nuclei (Q-S). Scale bar applies to all images. Images shown are representative median values in the graphs. * p<0.05; ** p<0.01; *** p<0.001.

Figure 4

MutantSOD1-induced motor neuron loss is ameliorated by steroid treatment in cranial motor nuclei but not the lumbar spinal cord. The total number of neurons (A, C, E, G), and the percentage of morphologically normal neurons (B, D, F, H) are reduced in all SOD1G93A groups compared with WT mice. The SOD1G93A 2B3-201 treatment group showed a general improvement in total number of neurons and the proportion of those with a normal morphological appearance, compared with both saline and free MP in the brainstem nuclei. No improvement was seen in the spinal cord. Frames I and J show an example of a neuron defined as morphologically normal (healthy, I) and an abnormal neuron with significant vacuolation (J). * p<0.01; ** p<0.001.

Figure 5

Iba1 positive microglia are not altered following steroid treatment. WT Iba1 immunoreactivity was very low in the cranial motor nuclei (A, E, I), and in the spinal cord (M). SOD1G93A mice treated with saline (B, F, J, N), free MP (C, G, K, O) and 2B3-201 (D, H, L, P) showed much stronger staining for Iba1, both in the three brainstem nuclei and the spinal cord. No difference was found in immunoreactivity between SOD1G93A treatment groups. Panels Q, S, U and W show quantification of Iba1 staining intensity in the brainstem motor nuclei and spinal cord, and panels R, T, V and X show the density of Iba1+ microglia in these same regions, showing increased expression in all SOD1G93A treatment groups compared with WT mice, but no group differences among treatments. Scale bar applies to all images. * p<0.05; ** p<0.01; *** p<0.001.

Figure 6

GFAP positive astrocytes reduced by steroid treatment. There is very little GFAP staining in the grey matter in WT mice, either in the cranial motor nuclei (A, E, I), or spinal cord (M). In the SOD1G93A saline group there was a marked GFAP up-regulation in the trigeminal (B), facial (F) and hypoglossal (J) nuclei, as well as the ventral spinal cord (N). A similar staining pattern is observed in SOD1G93A mice treated with free MP (C, G, K, O). GFAP staining in SOD1G93A 2B3-201 mice, however, was significantly reduced in all cranial motor nuclei (D, H, L), but not in the ventral spinal cord (P). Panels Q, S, U and W show quantification of GFAP staining intensity in the brainstem motor nuclei and spinal cord, and panels R, T, V and X show the density of GFAP+ astrocytes in these same regions. There is increased expression in all SOD1G93A treatment groups compared with WT mice, but SOD1G93A mice treated with 2B3-201 have less of a GFAP response compared with those treated with saline and free MP. Scale bar applies to all images. * p<0.05; ** p<0.01; *** p<0.001.